common/sieve: Support machines without MMU

Not all powerpc CPUs provide MMU support. Define vm_available() that is
true by default but archs can override it. Use this to run VM tests.

Reviewed-by:

common/sieve: Support machines without MMU

Not all powerpc CPUs provide MMU support. Define vm_available() that is
true by default but archs can override it. Use this to run VM tests.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Andrew Jones <andrew.jones@linux.dev>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Message-ID: <20240612052322.218726-6-npiggin@gmail.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

show more ...

lib/vmalloc: Let arch code pass a value to its setup_mmu() helper

Add an inner __setup_vm() that takes an opaque param and passes said
param along to setup_mmu(). x86 will use the param to configur

lib/vmalloc: Let arch code pass a value to its setup_mmu() helper

Add an inner __setup_vm() that takes an opaque param and passes said
param along to setup_mmu(). x86 will use the param to configure its page
tables for kernel vs. user so that tests that want to enable SMEP (fault
if kernel executes user page) can do so without resorting to hacks and
without breaking tests that need user pages, i.e. that run user code.

Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20210622210047.3691840-10-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib/alloc.h: remove align_min from struct alloc_ops

Remove align_min from struct alloc_ops, since it is no longer used.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Reviewed-by: Krish S

lib/alloc.h: remove align_min from struct alloc_ops

Remove align_min from struct alloc_ops, since it is no longer used.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Reviewed-by: Krish Sadhukhan <krish.sadhukhan@oracle.com>
Message-Id: <20210115123730.381612-7-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib/vmalloc: add some asserts and improvements

Add some asserts to make sure the state is consistent.

Simplify and improve the readability of vm_free.

If a NULL pointer is freed, no operation is p

lib/vmalloc: add some asserts and improvements

Add some asserts to make sure the state is consistent.

Simplify and improve the readability of vm_free.

If a NULL pointer is freed, no operation is performed.

Fixes: 3f6fee0d4da4 ("lib/vmalloc: vmalloc support for handling allocation metadata")

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Message-Id: <20210115123730.381612-4-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib/alloc: simplify free and malloc

Remove the size parameter from the various free functions

Since the backends can handle the allocation sizes on their own,
simplify the generic malloc wrappers.

lib/alloc: simplify free and malloc

Remove the size parameter from the various free functions

Since the backends can handle the allocation sizes on their own,
simplify the generic malloc wrappers.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Message-Id: <20201002154420.292134-6-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib/alloc_page: complete rewrite of the page allocator

This is a complete rewrite of the page allocator.

This will bring a few improvements:
* no need to specify the size when freeing
* allocate sm

lib/alloc_page: complete rewrite of the page allocator

This is a complete rewrite of the page allocator.

This will bring a few improvements:
* no need to specify the size when freeing
* allocate small areas with a large alignment without wasting memory
* ability to initialize and use multiple memory areas (e.g. DMA)
* more sanity checks

A few things have changed:
* initialization cannot be done with free_pages like before,
page_alloc_init_area has to be used instead

Arch-specific changes:
* s390x now uses the area below 2GiB for SMP lowcore initialization.

Details:
Each memory area has metadata at the very beginning. The metadata is a
byte array with one entry per usable page (so, excluding the metadata
itself). Each entry indicates if the page is special (unused for now),
if it is allocated, and the order of the block. Both free and allocated
pages are part of larger blocks.

Some more fixed size metadata is present in a fixed-size static array.
This metadata contains start and end page frame numbers, the pointer to
the metadata array, and the array of freelists. The array of freelists
has an entry for each possible order (indicated by the macro NLISTS,
defined as BITS_PER_LONG - PAGE_SHIFT).

On allocation, if the free list for the needed size is empty, larger
blocks are split. When a small allocation with a large alignment is
requested, an appropriately large block is split, to guarantee the
alignment.

When a block is freed, an attempt will be made to merge it into the
neighbour, iterating the process as long as possible.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Message-Id: <20201002154420.292134-5-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib/vmalloc: vmalloc support for handling allocation metadata

Add allocation metadata handling to vmalloc.

In upcoming patches, allocation metadata will have to be handled
directly bt the lower lev

lib/vmalloc: vmalloc support for handling allocation metadata

Add allocation metadata handling to vmalloc.

In upcoming patches, allocation metadata will have to be handled
directly bt the lower level allocators, and will not be handled by the
common wrapper.

In this patch, the number of allocated pages plus a magic value are
written immediately before the returned pointer. This means that multi
page allocations will allocate one extra page (which is not worse than
what the current allocator does).

For small allocations there is an optimization: the returned address is
intentionally not page-aligned. This signals that the allocation
spanned one page only. In this case the metadata is only the magic
value, and it is also saved immediately before the returned pointer.
Since the pointer does not point to the begininng of the page, there is
always space in the same page for the magic value.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Message-Id: <20201002154420.292134-3-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib/vmalloc: allow vm_memalign with alignment > PAGE_SIZE

Allow allocating aligned virtual memory with alignment larger than only
one page.

Add a check that the backing pages were actually allocate

lib/vmalloc: allow vm_memalign with alignment > PAGE_SIZE

Allow allocating aligned virtual memory with alignment larger than only
one page.

Add a check that the backing pages were actually allocated.

Export the alloc_vpages_aligned function to allow users to allocate
non-backed aligned virtual addresses.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Message-Id: <20200706164324.81123-5-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib/vmalloc: fix pages count local variable to be size_t

Since size is of type size_t, size >> PAGE_SHIFT might still be too big
for a normal unsigned int.

Signed-off-by: Claudio Imbrenda <imbrenda

lib/vmalloc: fix pages count local variable to be size_t

Since size is of type size_t, size >> PAGE_SHIFT might still be too big
for a normal unsigned int.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Jim Mattson <jmattson@google.com>
Message-Id: <20200706164324.81123-2-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib/vmalloc: add locking and a check for initialization

Make sure init_alloc_vpage is never called when vmalloc is in use.

Get both init_alloc_vpage and setup_vm to use the lock.

For setup_vm we o

lib/vmalloc: add locking and a check for initialization

Make sure init_alloc_vpage is never called when vmalloc is in use.

Get both init_alloc_vpage and setup_vm to use the lock.

For setup_vm we only check at the end because at least on some
architectures setup_mmu can call init_alloc_vpage, which would cause
a deadlock.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Message-Id: <20200622162141.279716-9-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib/vmalloc: fix potential race and non-standard pointer arithmetic

The pointer vfree_top should only be accessed with the lock held, so
make sure we return a local copy of the pointer taken safely

lib/vmalloc: fix potential race and non-standard pointer arithmetic

The pointer vfree_top should only be accessed with the lock held, so
make sure we return a local copy of the pointer taken safely inside the
lock.

Also avoid doing pointer arithmetic on void pointers. Gcc allows it but
it is ugly. Use uintptr_t for doing maths on the pointer.

This will also come useful in upcoming patches.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Message-Id: <20200622162141.279716-7-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib: use PAGE_ALIGN

Since now PAGE_ALIGN is available in all architectures, start using it in
common code to improve readability.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Message-Id

lib: use PAGE_ALIGN

Since now PAGE_ALIGN is available in all architectures, start using it in
common code to improve readability.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Message-Id: <20200622162141.279716-4-imbrenda@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

x86: allow using memory above 4 GiB

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

page_alloc: allow initialization before setup_vm call

Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>

vmalloc: convert to alloc_ops

The final step moves the vmalloc and vmap implementation to generic code,
rewriting vmalloc and vfree as an alloc_ops implementation that is
installed by setup_vm.

Thi

vmalloc: convert to alloc_ops

The final step moves the vmalloc and vmap implementation to generic code,
rewriting vmalloc and vfree as an alloc_ops implementation that is
installed by setup_vm.

This way all architectures can benefit from it just by calling setup_vm
and providing the implementation of install_page.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

x86: use alloc_phys

As the next step in making vmalloc/vfree available to every other architecture,
register the available memory with phys_alloc for x86 too, and add a service
to return the unused

x86: use alloc_phys

As the next step in making vmalloc/vfree available to every other architecture,
register the available memory with phys_alloc for x86 too, and add a service
to return the unused phys_alloc region. This makes setup_vm architecture-
independent.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

lib: start moving vmalloc to generic code

For now, vmalloc provides a primitive that allocates contiguous virtual
address. Together with a page allocator that allocates single
physical memory pages

lib: start moving vmalloc to generic code

For now, vmalloc provides a primitive that allocates contiguous virtual
address. Together with a page allocator that allocates single
physical memory pages, it will provide an implementation of
alloc_ops for when an MMU is enabled.

Before doing that, however, we need to move the page allocator and
give lib/alloc.c's malloc feature parity with lib/x86/vm.c's vmalloc.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...